What problem does it solve? Teams adopting AI coding agents on existing (brownfield) projects often don't know whether their current tech stack supports agent workflows well. This Skill evaluates the stack already in place against four quality gates (typed, convention-based, popular in training data, well-documented) and produces concrete compensation strategies instead of recommending a rewrite. ## Core Features & Use Cases - File-driven stack detection: Reads project markers like package.json, Cargo.toml, pyproject.toml, go.mod, and Gemfile to identify language, framework, build tool, test runner, package manager, CI/CD, and deployment target. - Quality gate scoring with evidence: Scores each component against the four gates, citing the specific file or config that proves each pass or fail, and outputs a scoring matrix. - Ready-to-paste compensation rules: For every failed gate, generates concrete AGENTS.md instruction entries (e.g., routing conventions, type annotation rules) so agents can work effectively despite gaps. - Use Case: A team with an Express.js codebase asks "is my stack agent-friendly?" The Skill detects the untyped, non-convention-based setup, scores it, and writes context/foundation/stack-assessment.md with paste-ready AGENTS.md rules for middleware order, validation at boundaries, and naming conventions. ## Quick Start Run the stack assessment on my current project directory and tell me how agent-friendly my stack is.
